Riverine and lake ecosystems are sources of global non-CO2 greenhouse gases (GHGs) and serve as sinks for pollutants, facing the dual challenges of mitigating GHG emissions and controlling water pollution. However, interactions between pollutant inputs and GHG emissions remain unclear. Herein, relying on a compiled data set of global non-CO2 GHGs and robust modeling, a watershed dissolved CH4 and N2O estimator is developed and validated on the global scale. Using the Dongting watershed (DTW) as a modeling example, various pollutant input scenarios were developed to explore the influence of changes in pollutant inputs on CO2-equivalent (CO2e) emissions from CH4 and N2O. Simulation results indicate that implementing pollutant inputs reduction measures in dissolved GHG hotspot areas will yield more efficient CO2e emission reduction benefits. Moreover, a critical paradox was revealed: while decreasing pollutant inputs leads to a sustained decline in direct CO2e (CO2eD) emissions, indirect CO2e (CO2eI) emissions from aquatic systems may show a minimal reduction in some cases. This paradox is closely tied to carbon-nitrogen ratio variations in aquatic system and can be well explained by carbon and nitrogen limitation principle, as defined by the Redfield ratio. Thus, our study suggests that cocontrol of carbon and nitrogen inputs within dissolved GHG hotspot areas is vital for achieving both water quality improvement and climate change mitigation simultaneously.